Vincetoxicum pycnostelma Kitag. (Apocynaceae - Asclepiadoideae) is a perennial plant that grows in grassland habitats of East Asia and western part of Amur (Kitamura et al. 1957). In Japan, most grassland habitat has been lost or reduced in size due to urbanization, land development, abandonment of traditional management of grasslands, or any combination of these factors. Thus, the abundance and density of V. pycnostelma have declined rapidly over the last several decades, and the extinction probability of this species in the next 100 years is estimated to be 96% (Environment Agency of Japan 2000). In 2007, this species was classified as “Near Threatened” in the Red Data List of Japan based on the IUCN criteria established by the Ministry of the Environment of Japan (2007).

Asclepiadoideae plants share unique flower morphologies and pollination systems (Figs. 1A, B). Pollen grains are contained in small sacs called pollinia, and 2 pollinia are joined to form a single pollinarium, which has a sticky appendage called corpusculum. In an individual flower, 5 stamens and 1 pistil are fused to a gynostegium, which has 5 narrow guide rails each leading to a stigmatic chamber and a corpusculum located above it. Pollination occurs in a 2-step process: (1) removal of a pollinarium occurs when a corpusculum in the guide rail attaches on the proboscis or a leg of an insect and is forcibly pulled from the flower, and (2) the insertion of the pollinarium is effected when it lodges in the stigmatic chamber (Wyatt 1978, Wyatt and Broyles 1994). Although the flower morphology of the asclepiadiodean plants is highly specialized, pollinators include a divergent group of insects including Lepidoptera, Diptera, Hymenoptera, and Coleoptera (Ollerton and Liede 1997, Ollerton et al. 2003).

View Figure

• Download as Powerpoint
• Download Figure

Knowledge of the pollination system is an essential component of a potentially successful conservation program for this endangered species. So far, there is only limited information on the pollination ecology of V. pycnostelma. It has the common floral characteristics of the Asclepiadoideae species described above (Konta et al. 1986), and only 7 insect species belonging to 4 families of 2 orders are recorded as pollinators (Yamashiro et al. 2008); however, effective pollinators have not been well defined or characterized. Asclepiadoideae plants also have a very low fruit set (Wyatt and Broyles 1990), which also has not been clearly characterized. The objectives of this study were to: (1) reveal the fauna visiting the flowers of V. pycnostelma, (2) identify the flower visitors that are effective pollinators, and (3) determine if the fruit set of V. pycnostelma is pollinator-limited.

Materials and Methods

A natural population of V. pycnostelma growing on the bank of the Kidzu River in Kamikoma (34°44′ 46″ N, 135°48′ 40″ E), Kyoto Prefecture, Japan, served as the site of this study. The population was comprised of approximately 500 V. pycnostelma plants.

Insects visiting flowers were counted and, when possible, collected on 28 June 2011 (1800 - 2100 h), 14 July 2011 (0,530 - 0,900 h), 22 July 2011 (1700 - 2130 h), 27 July 2011 (1700 - 2130 h), and 03 August 2011 (0,530 - 0,900 h). The ambient temperature at 30 cm above ground level was recorded every 30 min during each observation period. Collected insects were transported to the laboratory where they were identified to species using Inoue et al. (1982). Proboscis length (PL), wing expanse length (WEL), body length (BL), and head width (HW) of each specimen were measured using a micrometer caliper. Their bodies were examined under a stereomicroscope for the presence of pollinaria. Those with attached pollinaria were defined as effective pollinators.

At each observation, 5 flowers of V. pycnostelma were randomly collected and preserved in a fixed liquid of formalin, acetic acid, and 70% ethanol (5: 5: 90 in volume). The distance between corona lobes, the length from corona lobe to guide rail, and the width of the stigmatic chamber were measured using a micrometer caliper (Figs. 1C, D).

In addition, 124 flowers were randomly selected and removed from 25 plants on 14 July 2011 and transported to the laboratory. Each flower was examined under a stereomicroscope to count the number of pollinaria removed from the flower (Fig. 1C). The flower was then dissected, and the number of pollinaria in the stigmatic chambers was counted (Fig. 1D). The percentage of flowers from which pollinaria were removed, and the percentage of flowers with inserted pollinaria were calculated. Flowers with at least 1 inserted pollinarium were regarded as pollinated flowers.

From June to August 2011, 5 inflorescences after flowering, each having previously had 1 - 10 flowers, were examined from each of 20 randomly selected plants (639 total flowers). The number of fruits was counted, and fruit set percentage was calculated.

Results

Flower visitors and pollinators. At each observation, about 40 plants each with 1 or 2 inflorescences with 4 - 9 flowers each were monitored. Flowers were closed during most of the daylight hours with no flower-visiting insect activity observed whereas the flowers were closed. Sunset occurred at 1908 h on 22 July and at 1905 h on 27 July. Most of the flowers started to open at around 1730 h, and insects began to visit the flowers at 1800 h on 22 July and at 1900 h on 27 July. The maximum number of visitors observed was between 2000 and 2030 h. On 22 July and 27 July, 61 and 31 insects visited flowers between 1800 and 2130 h, respectively. Sunrise occurred at 0,453 h on 14 July and at 0,507 h on 3 August. Most of the flowers closed at around 0,700 h on both of those days. On 14 July and 3 August, 2 insects visited flowers between 0,530 and 0,700 h. All flower-visiting insects observed on 3 August were immediately captured by spiders, but such predatory activity was not observed in July.

Eighty-five insects visiting flowers were collected and identified; 81 of those were collected during the nights of 22 and 27 July. The 85 insects represented 27 species of 9 taxonomic families of 3 orders (Table 1). Of the 85 collected and identified, 80 were small adult lepidopterans, 4 were green lacewings (Neuroptera: Chrysopidae), and 1 was a crane fly (Diptera: Thipulidae). Fifteen insects (17.6%) had pollinaria of V. pycnostelma attached either to a proboscis (10.6%) or a leg (7.1%) (Fig. 2). All 15 were adult lepidopterans (14 species in families Arctiidae, Crambidae, Geometridae, Noctuidae, Pyralidae, and Tortricidae). The crane fly (Elephantomyia sp.) was captured on a flower with its proboscis in the guide rail of the flower; yet, no pollinaria were observed on the insect.

Table 1. List of insect species observed visiting the flowers of V. pycnostelma in Kyoto, Japan.

View Fullscreen

View Figure

• Download as Powerpoint
• Download Figure

Measurements of flower structures showed that the distance (mean ± SD) between adjacent corona lobes was 0.70 ± 0.07 mm, the distance between corona lobe and guide rail was 1.18 ± 0.86 mm, and the width of the stigmatic chamber was 0.71 ± 0.07 mm. Proboscis length of insects with attached pollinaria ranged from 0.9 - 4.5 mm (Table 2). Three of the insects with a proboscis ≥ 4.5 mm had pollinaria attached to their legs. Regression analysis failed to show any significant interaction of body length (< 13 mm) and wing expanse (< 23 mm) with presence or absence of pollinaria on the body.

Table 2. Mean and range of proboscis length, wing expanse, body length, and head width of insects visiting the flowers of V. pycnostelma in Kyoto, Japan.

View Fullscreen

Pollinaria removal, insertion and fruit set. Of the 620 pollinaria that should have been produced by the 124 flowers, 135 pollinaria (21.8%) had been removed before the time of sampling. The same number of flowers had 45 pollinaria (7.3%) inserted into their stigmatic chambers. The mean number of pollinaria removed per flower was 1.09, and the mean number of pollinaria inserted per flower was 0.36. Although 28.1% of the flowers received at least 1 pollinarium, the percentage fruit set of the flowers was as low as 2.3%.

Discussion

Yamashiro et al. (2008) reported that 17 species of insects representing 7 families and 2 orders visited the flowers of V. pycnostelma in Aichi and Miyagi Prefectures of Japan. Using results of a molecular phylogenetic and cladistics analysis, they further proposed that a clade of Vincetoxicum, including V. pycnostelma as well as the majority of other Japanese species of Tylophora-Vincetoxicum complex group, may have been derived from an ancestral state of dipteran pollination, adopting the observed moth pollination. However, V. pycnostelma was an exception in their study in that its most frequent pollinator was Limonia spp. (Diptera: Tipulidae).

In our study, all of the insects found to carry pollinaria of V. pycnostelma were small moths, suggesting that this species is not an outlier in the moth-pollinated clade. The nocturnal blooming of the flower of V. pycnostelma is surely one of the factors that facilitate pollination by nocturnal moths.

The moths carrying pollinaria of V. pycnostelma in our observations were all common species of rural environments in Japan. Host plants of their larvae are not known for all of them, but many are reported to be polyphagous (Inoue et al. 1982). The larvae of the most frequent pollinators, e.g., Rivula sericealis (Scopoli), feed on a range of common wild Gramineae. Thus, the pollinators of V. pycnostelma are specialized neither to the habitat of V. pycnostelma nor to V. pycnostelma itself. This suggests that the paucity of pollinator fauna is neither a major factor contributing the general decline of V. pycnostelma populations nor a factor to which priority should be focused in conservation plans for the species.

Almost all flower-visiting insects (except Tipulidae) had heads wider than 0.8 mm, which was wider than the distance between the corona lobes (0.7 mm) of the flowers. This indicates that those insects cannot insert their heads between corona lobes and, thus, likely approach the nectar source in the stigmatic chamber by extending and inserting their proboscis from outside of coronas. Thus, possession of a proboscis longer than the distance between the corona lobe and guide rail is needed to facilitate the feeding. This also is a likely reason why flower-visiting insects with a shorter proboscis < 0.9 mm in length did not have pollinaria on their proboscis. The distance between corona lobe and guide rail (1.18 mm) was nearly equal to the shortest proboscis of pollinators (0.9 mm), suggesting that insects with a proboscis shorter than the length from corona lobe to guide rail cannot insert pollinaria into the stigmatic chamber and, hence, cannot contribute to pollination.

Flower-visiting insects with proboscis longer than 4.5 mm did not have pollinaria attached to their proboscis. Although the reason for this is not clear, a possible explanation is that the longer proboscis also tends to be thicker and, thus, cannot be inserted in the opening of the guide rail of the V. pycnostelma flower. Instead, some of the insects with long (or thick proboscis) occasionally carried pollinaria on their legs. Attachment of pollinaria on insect legs is widely observed in Asclepiadoideae (Kunze and Liede 1991, Liede and Whitehead 1991) and can be one of the usual modes of pollination.

The number of removed pollinaria was 1.09 per flower and the number of inserted pollinaria was 0.36 per flower. These values are within the range reported on other Asclepiadoideae species (0.19 - 2.62 and 0.08 - 2.16, respectively) (Kunze and Liede 1991, Wolff et al. 2008) and suggests an efficient method of pollination by which Asclepiadoideae plants transfer their pollen to the stigma, if self-pollination (geitonogamy) is not a concern.

The low percentage of fruit set (2.3%) was also within the range (0.33 - 5.0%) reported on other Asclepiadoideae (Wyatt and Broyles 1990). This low fruit set is not likely due to quantitative pollen restriction because 28.2% of the flowers were found to have received at least 1 pollinarium. Possible alternative hypotheses may include self-incompatibility in that some Asclepiadaceae species possess postzygotic self-incompatibility (Lipow and Wyatt 2000). If V. pycnostelma also possess postzygotic self-incompatibility and most pollinaria inserted into stigmatic chamber came from the flower from the same plant, then fertilization did not occur and fruit was not set. Resource limitation also could be a contributing factor. Vincetoxicum pycnostelma produces big fruits compared with their plant size. They might have to limit resources allocated to sexual reproduction by mechanisms such as selective abortion as a means for compensating survival or growth of the parent plant. These should be also examined in further studies. Although these observations are limited to 1 site and to a short period of time, these results could be important for elucidating effective pollinators of V. pycnostelma.